latency-sensitive-bench/demon_attack_fix_latency_1

该数据集是一个用于机器学习任务的数据集，包含图像、文本提示和动作相关的数据，可能用于强化学习或交互式环境中的决策任务。特征包括图像、分割标识、提示文本、动作ID、动作文本、延迟时间、奖励值、完成标志、剧集索引、时间步、种子值和确定性标志。数据集分为训练集和验证集，训练集包含288,000个样本，验证集包含72,000个样本，总数据量约为320.23 MB。数据集旨在支持模型训练和评估，特别是在基于视觉和文本输入的决策场景中。This dataset is designed for machine learning tasks, containing image, text prompt, and action-related data, likely for decision-making in reinforcement learning or interactive environments. Features include image, split identifier, prompt text, action ID, action text, latency, reward value, done flag, episode index, time step, seed value, and deterministic flag. The dataset is split into training and validation sets, with 288,000 examples in the training set and 72,000 examples in the validation set, totaling approximately 320.23 MB. It aims to support model training and evaluation, particularly in scenarios involving visual and textual input for decision-making.